Concept: Administrative Health Data

Introduction

This concept contains a description of what Administrative Health data is; a list of some of the characteristics of administrative health data; a link to the MCHP Data Repository web page that lists the administrative health data sets in the MCHP Repository, and a list of MCHP publications that discuss aspects of administrative data validation.

What is Administrative Health Data ?

The collection of administrative health data was developed by provincial governments with the introduction of universal medical care insurance. The administrative health data housed in the Manitoba Population Research Data Repository at MCHP contains de-identified records for virtually all contacts with the provincial health care system. Examples of this include physician visits, hospitals discharge abstracts, personal care homes, home care, and pharmaceutical prescriptions.

Since the 1970s, Manitoba Health, the provincial agency that administers Manitoba's universal health insurance program, has provided copies of computerized health care utilization files to MCHP. The data derive from information contained in the Manitoba Health Services Insurance Plan Registry and from health insurance claims routinely filed by physicians and health care facilities with Manitoba Health. All individuals registered with the provincial health care system possess a 9-digit personal health identification number (PHIN), which exists only in encrypted form in the MCHP data bases. Utilization and follow-up information can be obtained for all Manitoba Health registrants.

Records may also be present for:


• Manitoba registrants admitted to a facility in a non-reciprocating province or outside Canada;
  
• Manitoba residents not yet eligible for coverage (payment is reciprocal);
  
• Ex-Manitoba residents within three months of leaving for a reciprocating province;
  
• Non-Manitobans admitted to Manitoba facilities

The improvement in computer and information technology has enabled the expanded use of administrative health data for epidemiologic and evaluative research. Administrative health data can be used to study the epidemiology of disease, patterns of physician practice, and the relationship between health care and outcomes (Spasoff, 1999). Moves to constrain health care costs and restructure health services have resulted in the utilization of administrative data for the evaluation of the efficiency and effectiveness of services.

Characteristics of Administrative Health Data

Some of the characteristics of administrative health data include:


• Population-Based: The entire population of the province is covered by the Manitoba Health Services Insurance Plan. Non-participation is minimal since Manitoba residents do not pay premiums to register for insured benefits.
  
  
• Unique Identifier: Use of a consistent set of identifiers (with identification numbers of both patients and physicians encrypted to ensure confidentiality) permits researchers to build histories of individuals across files. For example, individuals who are discharged from hospital can be linked to the medical claims file in order to determine if adverse events are being treated in physicians' offices.
  
  
• Follow-Up: Migration into and out of the province, and mortality can be traced from 1970 onward. Researchers can track groups of subjects through time in order to determine if individuals receiving a given intervention truly have no adverse outcomes, or if no adverse events appear to be occurring because the individual has left the province or died.
  
  
• Longitudinal: Multiple years of data permit studying change over time for numerous variables (e.g.: going back to 1970 for hospital and physician claims). Such extensive historical data is important for studying the determinants of health. Since patient identifiers, physician and hospital numbers have been standardized across all files, histories can be developed, for example, of individuals' contact with physicians or hospitals over time, of all care rendered by physicians over time, or of hospital or PCH admissions over time.
  
  
• Secondary Data Sources: Because administrative health databases were designed for purposes other than research, care is required to ensure, for example, accurate counts of individuals having multiple sites of residence or having procedures done out of province. Records only include information for individuals who use the services during the period of interest. Potential limitations include clinically imprecise coding, absence of key clinical data on processes and outcomes, and inconsistent recording of provider information (Huston and Naylor, 1996). ( [Abstract on PubMed] )
  
  
• Reliability and Validity: Several studies have found the administrative health data to have a high degree of reliability and validity. See the Administrative Data Validation Studies section below for more information.

Administrative Health Data in the Manitoba Population Research Data Repository

For a list of administrative health data sets in the Manitoba Population Research Data Repository, please click on the following link:


• Administrative Health Data.

Administrative Data Validation Studies

The following list identifies studies that have investigated data validation with administrative data:


• Black C, Roos NP: Administrative data: Baby or bathwater? (Editorial). Med Care 1998; 36(1):3-5 [abstract ; MCHP view] .
  
• Robinson JR, Young TK, Roos LL, Gelskey DE: Estimating the burden of disease: comparing administrative data and self-reports. Med Care 1997 35: 932-47 [abstract ; MCHP view] .
  
• Roos LL, Gupta S, Soodeen RA, Jebamani L: Data quality in an information-rich environment: Canada as an example. Can J Aging 2005;24 Suppl 1(Suppl 1):153-170 [abstract ; MCHP view] .
  
• Roos LL, Nicol JP: A research registry: uses, development, and accuracy. J Clin Epidemiol 1999;52(1):39-47 [abstract ; MCHP view] .
  
• Roos LL, Nicol JP, Cageorge SM: Using administrative data for longitudinal research - comparisons with primary data collection. J Chron Dis 1987; 40(1):41-49 [abstract ; MCHP view] .
  
• Roos LL, Romano PS, Fergusson P: Administrative data bases. P Armitage & T Colton (eds), Encyclopedia of Biostatistics. Chichester, England: John Wiley and Sons; 1998:62-73 [MCHP view] .
  
• Roos LL, Roos NP, Cageorge SM, Nicol JP. How good are the data: Reliability of one health care data bank. Med Care 1982, 20: 266-276 [abstract ; MCHP view]
  
• Roos LL, Sharp SM, Wajda A: Assessing data quality - a computerized approach. Soc Sci Med 1989 ; 28(2):175-182.
  
• Roos LL, Stranc L, James RC, Li JW: Complications, comorbidities, and mortality: Improving classification and prediction. Health Services Research 1997; 32(2):229-238. [abstract ; MCHP view]
  
• Roos LL, Walld R, Wajda A, Bond R, Hartford K: Record linkage strategies, outpatient procedures and administrative data. Med Care 1996;34(6):570-582 [abstract ; MCHP view] .
  
• Roos NP, Shapiro E. Medical Care: Health and Health Care: Experience with a Population-Based Health Information System. Med Care 1995;33(12 Suppl):DS1-DS146 [MCHP view] .
  
• Wray NP, Ashton CM, Kuykendall DH, Hollingsworth JC: Using administrative databases to evaluate the quality of medical care: a conceptual framework [Review]. Soc Sci Med 1995; 40(12):1707-1715 [abstract] .

Note: Other validation studies have been carried out on specific data sets in the Manitoba Population Research Data Repository. These studies are listed in the Data Descriptions for a specific data set.